Prediction of students’ feedback on faculty performance using stacking ensemble method: Machine learning algorithm

Abstract

Students’ feedback is fundamental for the growth and development of higher education institutions. Feedback and comments from students are an extremely useful and valuable source of information that reflects the quality of education or educational services received by students. However, the effective management of qualitative opinions of students is a challenge. Undeniably, many organisations deal with quantitative feedback effectively, while qualitative feedback is either manually processed or ignored. This paper proposes an opinion mining or sentiment analysis system using a stacking ensemble-based method. Furthermore, four base models, comprising various base-level classifiers, including logistic regression (LR), support vector machine (SVM), multilayer perceptron (MLP), and Naïve Bayes (NB), predict the orientations as positive, negative, or neutral. The system has been evaluated using performance metrics such as accuracy, precision, recall and F1-measure; and compared with similar models. Experimental results show that the four base-independent algorithms yield the following classification accuracies: LR algorithm, 79.05%; SVM, 81.76%; MLP, 50.68%; and Multinomial Naïve Bayes, 50.68%. These forecasts can be used by Nigerian Public universities and higher education institutions to improve the educational system and assist students to receive a better and quality education.